Vascular risk burden is a key player in the early progression of Alzheimer’s disease

1. Introduction

A new era of trials in individuals with preclinical Alzheimer’s disease is starting given the assumption that better outcomes could be achieved with interventions performed before the presence of extensive damage and cognitive symptoms.^{1, 2} The preclinical stage of Alzheimer’s disease has been characterized as the presence of biomarker evidence of amyloid-β (Aβ) and tau pathologies in cognitively unimpaired (CU) individuals.³ Subjects in this stage are at higher risk for Alzheimer’s disease clinical progression; however, many of them never progress to cognitive impairment, suggesting that other simultaneous pathological processes are involved.⁴ Therefore, it is important to understand the additional factors contributing to Alzheimer’s disease progression for the development of effective therapeutic strategies.

Vascular risk factors (VRFs), such as hypertension, diabetes mellitus, smoking, and hypercholesterolemia, are well-established risk factors for developing Alzheimer’s disease dementia.^5–11 These conditions are associated with cerebrovascular lesions in neuropathologically-confirmed Alzheimer’s disease patients,^{12, 13} and the presence of these brain injuries contributes to dementia onset.^12–15 Furthermore, as multiple VRFs often coexist and gradually increase Alzheimer’s disease risk,⁶ recent research focused on these conditions in combination (i.e., burden) rather than individually. Nevertheless, it remains to be elucidated whether vascular risk and Alzheimer’s disease pathophysiology have additive or synergistic effects on neurodegeneration and cognitive decline. Also, the direct effect of VRFs on Alzheimer’s disease pathophysiology is still not completely understood. While some studies support a possible relation with Aβ or tau deposition,^16–19 others point to the opposite direction.^{12, 20–23}

Determining whether vascular risk factors synergistically potentiate Alzheimer’s disease progression has potential implications for clinical trials. It could suggest that the combination of therapies targeting Aβ and tau pathologies, as well as VRFs, may enhance treatment outcomes. With this in mind, the main goal of the present study was to investigate whether VRF burden synergistically interacts with Alzheimer’s disease pathophysiology to accelerate neurodegeneration and cognitive decline in CU individuals. Secondarily, we also aimed to assess whether VRF burden is related to changes in Aβ and tau biomarkers over time.

2. Materials and methods

Data used in the preparation of this article were obtained from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database (adni.loni.usc.edu; last accessed May 2021). The ADNI was launched in 2003 as a public-private partnership, led by Principal Investigator Michael W. Weiner, MD. The primary goal of ADNI has been to test whether serial magnetic resonance imaging (MRI), positron emission tomography (PET), other biological markers, and clinical and neuropsychological assessment can be combined to measure the progression of mild cognitive impairment (MCI) and early Alzheimer’s disease. Detailed information concerning inclusion and exclusion criteria has already been described.²⁴ Of note, participants were recruited between the ages of 55 and 90 years, completed at least 6 years of education, were fluent in Spanish or English, had a Hachinski ischemic score less than or equal to four, and had screening/baseline MRI scans without evidence of infection, infarction, or other focal lesions (subjects with multiple lacunes or lacunes in a critical memory structure were excluded). Institutional Review Boards of all involved sites approved the ADNI study and all research participants or their authorized representatives provided written informed consent.

3. Results

A total of 503 participants (43.7% males) were assessed in this study, of whom 13.1% had CSF biomarker evidence of preclinical Alzheimer’s disease [(AT)+], and 45.9% presented an elevated VRF burden (V+). Sample demographics, biomarker and clinical characteristics are summarized in Table 1. In addition, detailed demographic information regarding subsamples used to evaluate the trajectory of plasma NfL and CSF biomarkers - Aβ_1-42 and p-tau₁₈₁ – are available in the Supplementary Table 2 and Supplementary Table 3. Regarding model inspections, multicollinearity was not identified in any LME model used. Besides, survival model diagnostics indicated that proportional hazards assumption was not violated.

3.1. VRF burden and Alzheimer’s disease pathophysiology act synergistically on plasma NfL levels

LME model coefficients for the associations between VRF burden, Alzheimer’s disease pathophysiology, and longitudinal plasma NfL can be found in Table 2, Model A. At baseline, preclinical Alzheimer’s disease status was marginally associated with higher concentrations of plasma NfL (β = 6.23, P = 0.082). On the other hand, VRF burden was not significantly associated with baseline plasma NfL levels (P = 0.542), nor was the interaction between VRF burden and Alzheimer’s disease pathophysiology (P = 0.754). Concerning plasma NfL longitudinal trajectory, there was a significant three-way interaction (VRF burden x Alzheimer’s disease pathophysiology x time; β = 5.08, P = 0.016), indicating that an elevated VRF burden acted synergistically with preclinical Alzheimer’s disease to increase plasma NfL concentrations longitudinally. To confirm the presence of a synergistic interaction, we found that the interaction effect of VRF burden and Alzheimer’s disease pathophysiology was greater than the sum of their independent effects (Table 3, Model A). Interestingly, VRF burden x time and Alzheimer’s disease pathophysiology x time interaction terms were not significant (P = 0.351 and P = 0.793, respectively). This shows that VRF burden and Alzheimer’s disease pathophysiology did not associate independently with changes in NfL levels over time. For results stratified by groups, see Fig. 1A.

View this table:

• View inline
• View popup
• Download powerpoint

Table 2. Liner mixed-effects model coefficients.

View this table:

• View inline
• View popup
• Download powerpoint

Table 3. Independent and interactive effects of vascular risk factor (VRF) burden and Alzheimer’s disease pathophysiology on longitudinal plasma neurofilament light (NfL) and cognitive trajectories.

• Download figure
• Open in new tab

Figure 1. Elevated vascular risk factor (VRF) burden accelerates neurodegeneration and cognitive decline in subjects with preclinical Alzheimer’s disease.

Mean predicted trajectories and 95% confidence intervals estimated from linear mixed-effects models according to baseline and VRF burden status and Alzheimer’s disease pathophysiology status. (A) Longitudinal neurodegeneration measured by plasma neurofilament light (NfL) in a 4-year follow-up period and (B) longitudinal cognitive trajectory indexed by the modified Preclinical Alzheimer’s Cognitive Composite (mPACC) in a 6-year follow-up period. Each model was adjusted for age, sex, years of education, apolipoprotein E ε4 status, and their interaction with time.

3.3. VRF burden is not associated with Alzheimer’s disease CSF biomarkers

In LME models evaluating CSF Aβ_1-42 and p-tau₁₈₁ trajectories, neither the main effect of VRF burden nor the VRF burden x time interaction was significant (Fig. 2A and Model C in Table 2 for CSF Aβ_1-42 and Fig. 2B and Model D in Table 2 for p-tau₁₈₁; all P ≥ 0.314). Hence, VRF burden was not associated with baseline CSF Aβ_1-42 and p-tau₁₈₁ levels nor with changes in its levels over time.

• Download figure
• Open in new tab

Figure 2. Vascular risk factor (VRF) burden is not associated with changes over time in cerebrospinal fluid (CSF) Aβ_1-42 and p-tau₁₈₁ levels.

Mean predicted trajectories and 95% confidence intervals estimated from linear mixed-effects models according to baseline VRF burden status. (A) CSF Aβ_1-42 longitudinal trajectory in a 6-year follow-up and (B) CSF p-tau₁₈₁ longitudinal trajectory in a 6-year follow-up period. Each model was adjusted for age, sex, years of education, apolipoprotein E ε4 status, and their interaction with time. Aβ_1-42 = amyloid-β_1-42; p-tau₁₈₁ = tau phosphorylated at threonine 181.

3.4. Increased risk of clinical progression in (AT)+V+ individuals

Kaplan-Meier curves demonstrate that the (AT)+V+ group presented a well-separated survival probability curve in a 6-year follow-up period, while the trajectories of the remaining groups [(AT)-V-, (AT)-V+, and (AT)+V-] were overlapping among themselves (Fig. 3A). The log-rank test showed a significant difference in estimated survival probabilities across the four groups (P < 0.0001). Cox-proportional hazards models corroborated that only the (AT)+V+ group had a significantly higher risk of clinical progression compared to the reference group [(AT)-V-] in 6 years (adjusted HR = 3.5, P < 0.001; Table 4). Post-hoc comparisons demonstrated that (AT)+V+ group presented a significantly higher risk of clinical progression in comparison to (AT)-V- (P = 0.005) and (AT)-V+ (P = 0.012) groups, whereas it did not reach statistical significance in comparison to the (AT)+V-group (P = 0.127). Additionally, no differences were observed in pairwise contrast analysis among the (AT)-V-, (AT)-V+, and (AT)+V-groups. In further analyses not adjusting for potential confounders, we observed similar findings (Supplementary Table 4). To better understand the unexpected slow conversion in the (AT)+V-group, we conducted a few exploratory analyses testing clinical deterioration detected by the CDR-SB scale in the whole sample (Fig. 3B) and in individuals under 75 years (Fig. 3C), as well as survival probabilities using WMH status (instead of VRF burden status) for clinical conversion (Fig. 3D) and for changes in the CDR-SB scale (Fig. 3E) in the whole sample. The reduced number of conversions is a limiting factor that does not allow definitive conclusions. This issue is properly discussed in the limitation section and deserves further investigation.

View this table:

• View inline
• View popup
• Download powerpoint

Table 4. Adjusted Hazard Ratios of clinical progression according to vascular risk factor (VRF) burden and Alzheimer’s disease pathophysiology status.

• Download figure
• Open in new tab

Figure 3. Individuals with preclinical Alzheimer’s disease and elevated vascular risk factor (VRF) burden have a greater risk of clinical progression.

(A) Conversion to cognitive impairment [i.e., mild cognitive impairment (MCI) or dementia] in the whole sample according to baseline VRF burden status and Alzheimer’s disease pathophysiology status. (B) Changes in the clinical dementia rating scale sum of boxes (CDR-SB) scale in the whole sample according to baseline VRF burden status and Alzheimer’s disease pathophysiology status. (C) Changes in the CDR-SB scale in individuals under 75 years according to baseline VRF burden status and Alzheimer’s disease pathophysiology status. (D) Conversion to cognitive impairment (i.e., MCI or dementia) in the whole sample according to baseline white matter hyperintensities (WMH) status and Alzheimer’s disease pathophysiology status. (E) Changes in the CDR-SB scale in the whole sample according to baseline WMH status and Alzheimer’s disease pathophysiology status. Kaplan-Meier curves with 95% confidence intervals display survival probability for each group. Significance level obtained from the log-rank test is also reported.

4. Discussion

In the present study, we showed that VRF burden interacts synergistically with Alzheimer’s disease pathophysiology to drive longitudinal increases in plasma NfL levels, as well as longitudinal decline in mPACC scores in CU individuals. Nevertheless, VRF burden alone was not associated with changes in CSF Alzheimer’s disease biomarkers (CSF Aβ_1-42 and p-tau₁₈₁). Additionally, we demonstrated that the presence of both elevated VRF burden and preclinical Alzheimer’s disease in CU individuals is associated with higher progression rates to MCI or dementia. These findings suggest that the impact of VRF burden on neurodegeneration and cognitive decline varies according to Alzheimer’s disease pathophysiology status in CU subjects.

Here, we observed that VRF burden and Alzheimer’s disease pathophysiology were synergistically associated with longitudinal neurodegeneration measured by plasma NfL. Even though it focused on another neurodegeneration biomarker, a previous study reported that CU individuals with an elevated Framingham Risk Score presented higher rates of increase in CSF t-tau levels when abnormal to Aβ and tau biomarkers at baseline.⁴³ Given that the association was restricted to the A+T+ group, this finding further supports the notion of an interactive association between vascular risk and Alzheimer’s disease pathophysiology. Also, our results are in agreement with cross-sectional neuroimaging evidence that vascular risk and Aβ burden were interactively associated with lower cortical thickness in Alzheimer’s disease vulnerable (posterior) brain regions in CU and mildly impaired subjects.⁴⁴ Taken together, this complimentary evidence highlights the importance of considering the VRF burden when studying the processes underlying brain damage in preclinical Alzheimer’s disease.

The notion that VRF burden and Alzheimer’s disease pathophysiology interact to promote cognitive impairment is also supported by recently published evidence in the literature. In participants from the Harvard Aging Brain Study (HABS), it was observed a synergistic association of vascular risk and Aβ burden with prospective cognitive decline in CU elderlies.²² Similarly, another longitudinal study including CU older adults demonstrated that an increased Framingham Risk Score was associated with higher rates of cognitive decline only in the A+T+ group, indicating an interactive effect.⁴³ Notwithstanding, divergent results have also been reported. A recent investigation evaluating CU participants from the Biomarkers for Older Controls at Risk for Dementia (BIOCARD) study concluded that midlife vascular risk and Alzheimer’s disease pathophysiology – defined by CSF biomarkers – presented additive rather than synergistic effects on cognitive decline over a mean follow-up of 13.9 years.²⁰ Although assessing VRFs in midlife and having a longer follow-up is a clear strength of this work, other factors could account for the divergent results, such as vascular risk assessment (dichotomization by 0 or ≥ 1 evaluating the following conditions: hypertension, hypercholesterolemia, diabetes, smoking, and obesity) and cutpoints used for Alzheimer’s disease biomarkers (based on tertiles calculated considering midlife Alzheimer’s disease biomarker concentrations). Together, these findings reinforce that, even without a clear consensus in the literature, VRFs and Alzheimer’s disease pathophysiology often coexist and play pivotal roles in brain aging.

Furthermore, our results suggest that VRF burden does not directly impact Alzheimer’s disease pathophysiology. Previous studies have reported inconsistent results concerning the links between VRFs and Alzheimer’s disease pathophysiology. While some investigations found that VRFs were cross-sectionally associated either with higher Aβ or tau burden,^{18, 19} most evidence is against the presence of such relations.^{12, 22, 23} Nonetheless, contrary to our results, a recent study showed that the cumulative number of midlife VRFs was associated with late-life Aβ deposition,¹⁶ as well as that high vascular risk and elevated Aβ burden at baseline were interactively associated with increased tau PET signal in the inferior temporal cortex measured nearly three years later.¹⁷ On the other hand, our finding that the VRF burden does not present short-term influence on Alzheimer’s disease pathophysiology is in line with recent studies evaluating the associations between composite vascular risk scores – calculated based on the cumulative number of vascular risk factors or the Framingham Risk Score – and longitudinal changes in the CSF Aβ_1-42 and p-tau₁₈₁ levels.^{20, 21, 43} To rule out or confirm a possible association between VRF burden and Alzheimer’s disease pathophysiology, further cohort studies with longer follow-up and increased sample size, ideally using neuroimaging biomarkers, are needed. Lastly, it is vital to acknowledge the possibility of differential mechanisms and weights specific for each VRF, which was beyond the scope of the present investigation.

Not rarely do individuals without cognitive impairment present Alzheimer’s disease pathophysiology,^45–49 highlighting the role of both resilience mechanisms and concomitant pathological processes in the clinical expression of Alzheimer’s disease. In this context, it has been proposed that vascular dysfunction has an early role in Alzheimer’s disease progression.⁵⁰ There are different potential mechanisms by which vascular factors contribute to cognitive impairment and dementia, such as reduction in cerebral blood flow and hypoxia, blood-brain barrier (BBB) breakdown, endothelial dysfunction, systemic inflammation and oxidative stress, and disruption of trophic coupling.⁵¹ Our results support the notion that CU individuals exposed to higher VRFs might have a decreased threshold for cognitive decline and neurodegeneration induced by Alzheimer’s disease pathophysiology. As previously suggested, a possible explanation is that VRFs influence the progression of Alzheimer’s disease through the promotion of cerebrovascular injuries rather than through a direct effect on Alzheimer’s disease pathophysiology.²³ Additionally, BBB breakdown, an important feature in early Alzheimer’s disease^52–55 and a potential biomarker of cognitive dysfunction,⁵⁶ is associated with both Alzheimer’s disease pathophysiology and VRFs, but at different molecular levels, stressing the role of brain vasculature in cognitive impairment. Hence, VRFs appear to impact brain resilience mechanisms against the deposition of Aβ-containing extracellular neuritic plaques and tau-containing neurofibrillary tangles.

At the moment, there is no approved pharmacological treatment that can unquestionably stop or delay Alzheimer’s disease clinical deterioration. Since Alzheimer’s disease is a multifactorial disease, it is reasonable to consider that an effective therapy would need to have multiple targets, not only Aβ and tau accumulation. Also, given that Alzheimer’s disease pathophysiology starts to accumulate many years before the onset of clinical symptoms,⁵⁷ new clinical trials often focus on asymptomatic individuals presenting biomarker evidence of Aβ and tau pathologies (i.e., preclinical stages of Alzheimer’s disease).^{1, 2} The findings from the present work corroborate that the development of therapies targeting both Alzheimer’s disease pathophysiology and VRFs in Alzheimer’s disease preclinical stages could potentiate treatment response.

Similar to ADNI’s enrollment criteria, clinical trials for Alzheimer’s disease usually do not include individuals with an elevated cerebrovascular burden. To this end, enrollment criteria commonly exclude subjects with Hachinski ischemic scores equal or greater than four and having brain MRI lesions.^58–60 Excluding patients with significant cerebrovascular disease could be seen as a limitation of the present study as it makes the cohort underrepresented in terms of vascular pathology. However, we could still observe an interactive association of VRF burden with Alzheimer’s disease pathophysiology on longitudinal neurodegeneration and cognitive decline in CU individuals. These results suggest that future clinical trials in individuals with preclinical Alzheimer’s disease should not exclude a possible vascular contribution to Alzheimer’s disease progression by merely using the aforementioned enrollment criteria. On the other hand, a potential approach for studies aiming to solely focus on Alzheimer’s disease pathophysiology is to further apply the composite VRF burden score used in the present work.

Even though cognition is the primary outcome of interest in disease-modifying drug trials, individuals with preclinical Alzheimer’s disease can stay cognitively stable over many years.⁴ Therefore, an important limitation for the performance of these trials is the need for high sample sizes and extended follow-ups. In this context, the use of surrogate markers of disease progression can be a useful alternative. Besides being noninvasive and accessible, plasma NfL has been shown to be a robust neurodegeneration marker to monitor Alzheimer’s disease progression.^{31, 61} Nonetheless, to the best of our knowledge, no previous study has longitudinally appraised the associations of VRF burden and Alzheimer’s disease pathophysiology with plasma NfL trajectories. Similar to the pattern observed for cognition, our results showed that the synergy between VRF burden and Alzheimer’s disease pathophysiology led to longitudinal increases in plasma NfL. Together, these findings suggest that plasma NfL may be used as a surrogate to track therapeutic response in trials targeting vascular risk factors and Alzheimer’s disease pathophysiology.

From the perspective of preventive strategies, a recent report estimated that modifiable risk factors – including VRFs such as hypertension, smoking, and diabetes – are associated with nearly 40% of dementia cases, which could be potentially prevented or delayed.⁶² In fact, a recent study has suggested that the reduction in dementia incidence in participants of the Framingham Heart Study over the past three decades could be, at least in part, a consequence of improved management of VRFs.⁶³ Additionally, the multicenter Finnish Geriatric Intervention Study to Prevent Cognitive Impairment and Disability (FINGER) trial has shown that a multidomain lifestyle intervention – which involved controlling VRFs – provided benefits for cognitive performance in a population of older subjects at increased risk for cognitive decline and dementia.⁶⁴ In accordance with the literature evidence, the data presented in this investigation reinforces the importance of developing preventive strategies for reducing dementia societal burden, which must include maintaining a good vascular health.

Some limitations need to be highlighted to appropriately interpret our results. First, the ADNI study involves a selective population of highly educated and mostly white volunteers, which might further jeopardize the generalizability of our results, especially because different ethnic groups are not homogenously affected by VRFs.^65–67 Second, it is also necessary to consider the limited sample size, mainly of participants with preclinical Alzheimer’s disease. This issue is probably the reason for the wide confidence intervals in time-to-event analysis. Third, we did not consider the age of onset of the VRFs. Some VRFs may have an age-dependent effect – such as midlife and late-life exposures being differentially related to the risk of dementia and specifically Alzheimer’s disease.^{11, 68–71} Thus, we cannot determine whether our findings would differ considering only exposures that started in midlife or late-life. Fourth, the presence of VRFs was determined based on previous diagnosis and use of medications information collected in clinical interviews rather than diagnosis performed at study entry with objective measurements, potentially being a source of bias. Fifth, previous studies demonstrated that CSF Aβ_1-42/1-40 ratio better predict abnormal brain Aβ accumulation in comparison to CSF Aβ_1-42.⁷² However, it was not possible to determine Aβ positivity using the CSF Aβ_1-42/1-40 ratio in the present study due to limited CSF Aβ_1-40 data available.

In conclusion, we demonstrated that VRF burden and biomarker evidence of Alzheimer’s disease pathophysiology are synergistically associated with neurodegeneration and cognitive deterioration in CU individuals, whereas VRF burden does not influence Aβ and tau pathologies. Our results provide additional evidence for the performance of clinical trials targeting VRFs and Alzheimer’s disease pathophysiology. Additionally, these trials could have advantages from using plasma NfL as a surrogate to track therapeutic response. Importantly, further community-based studies with longer follow-up and bigger sample sizes are needed to confirm our findings.

Funding

J.P.F-S. receives financial support from CAPES (#88887.627297/2021-00). W.S.B is supported by CAPES (#88887.372371/2019-00 and #88887.596742/2020-00). B.B. receives financial support from CAPES (#88887.336490/2019-00). D.T.L. is supported by a NARSAD Young Investigator Grant from the Brain & Behavior Research Foundation (#29486). A.B. receives financial support from CAPES (#88882.345554/2019-01). C.T. receives funding from Faculty of Medicine McGill and IPN McGill. M.A.D.B. receives financial support from CNPq (#150293/2019-4). G.P. receives financial support from CAPES (#88882.345577/2019-01). A.L.B. is supported by the Swedish Alzheimer Foundation, Stiftelsen för Gamla Tjänarinnor, and Stohne Stiftelsen. J.T is supported by the Canadian Institutes of Health Research & McGill Healthy Brains Healthy Lives initiative. H.Z. is a Wallenberg Scholar supported by grants from the Swedish Research Council (#2018-02532), the European Research Council (#681712), Swedish State Support for Clinical Research (#ALFGBG-720931), the Alzheimer Drug Discovery Foundation (#201809-2016862), the AD Strategic Fund and the Alzheimer’s Association (#ADSF-21-831376-C, #ADSF-21-831381-C, and #ADSF-21-831377-C), the Olav Thon Foundation, the Erling-Persson Family Foundation, Stiftelsen för Gamla Tjänarinnor, Hjärnfonden (#FO2019-0228), the European Union’s Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie grant agreement No 860197 (MIRIADE), and the UK Dementia Research Institute at UCL. K.B. is supported by the Swedish Research Council (#2017-00915), the Alzheimer Drug Discovery Foundation (#RDAPB-201809-2016615), the Swedish Alzheimer Foundation (#AF-742881), Hjärnfonden (#FO2017-0243), the Swedish state under the agreement between the Swedish government and the County Councils, the ALF-agreement (#ALFGBG-715986), the European Union Joint Program for Neurodegenerative Disorders (#JPND2019-466-236), the NIH (#1R01AG068398-01), and the Alzheimer’s Association 2021 Zenith Award (#ZEN-21-848495). S.O.M. is supported by the Research Grant for Resilient Study (DECIT/CNPq #401821/2015-3), for Trident Study (PROADI-SUS Hospital Moinhos de Vento #NUP 25000.209767/2018-61), for Polypill and Stroke Riskometer for Prevention of Stroke and Cognitive Decline (PROADI-SUS HMV #NUP 25000.013007/2021-55), for Resilient Direct TNK and Extend IV (PROADI-SUS HMV #NUP 25000.0334940/2021-66). D.O.S. is supported by CAPES (#88887.185806/2018-00, #88887.507218/2020-00, and #88887.507161/ 2020-00), CNPQ/INCT (#465671/2014-4), CNPQ/ZIKA (#440763/ 2016-9), CNPQ/FAPERGS/PRONEX (#16/2551-0000475-7), and FAPERGS (#19/2551-0000700-0). P.R-N. is supported by the Canadian Institutes of Health Research (#MOP-11-51-31), the Alzheimer’s Association (#NIRG-12-92090 and #NIRP-12-259245), Fonds de Recherche du Québec - Santé (Chercheur Boursier; #2020-VICO-279314). T.K. is funded by the Swedish Research Council’s career establishment fellowship (#2021-03244), the Alzheimer’s Association Research Fellowship (#850325), the BrightFocus Foundation (#A2020812F), the International Society for Neurochemistry’s Career Development Grant, the Swedish Alzheimer Foundation (Alzheimerfonden; #AF-930627), the Swedish Brain Foundation (Hjärnfonden; #FO2020-0240), the Swedish Dementia Foundation (Demensförbundet), the Swedish Parkinson Foundation (Parkinsonfonden), Gamla Tjänarinnor Foundation, the Aina (Ann) Wallströms and Mary-Ann Sjöbloms Foundation, the Agneta Prytz-Folkes & Gösta Folkes Foundation (#2020-00124), the Gun and Bertil Stohnes Foundation, and the Anna Lisa and Brother Björnsson’s Foundation. T.A.P. is supported by the NIH (#R01AG075336 and #R01AG073267) and the Alzheimer’s Association (#AACSF-20-648075). E.R.Z. receives financial support from CNPq (#435642/2018-9 and #312410/2018-2), Instituto Serrapilheira (#Serra-1912-31365), Brazilian National Institute of Science and Technology in Excitotoxicity and Neuroprotection (#465671/2014-4), FAPERGS/MS/CNPq/SESRS–PPSUS (#30786.434.24734.231120170), and ARD/FAPERGS (#54392.632.30451.05032021).

Competing interests

H.Z. has served at scientific advisory boards and/or as a consultant for Abbvie, Alector, Annexon, AZTherapies, CogRx, Denali, Eisai, Nervgen, Pinteon Therapeutics, Red Abbey Labs, Passage Bio, Roche, Samumed, Siemens Healthineers, Triplet Therapeutics, and Wave, has given lectures in symposia sponsored by Cellectricon, Fujirebio, Alzecure and Biogen, and is a co-founder of Brain Biomarker Solutions in Gothenburg AB (BBS), which is a part of the GU Ventures Incubator Program. K.B. has served as a consultant, at advisory boards, or at data monitoring committees for Abcam, Axon, Biogen, JOMDD/Shimadzu. Julius Clinical, Lilly, MagQu, Novartis, Prothena, Roche Diagnostics, and Siemens Healthineers, and is a co-founder of Brain Biomarker Solutions in Gothenburg AB (BBS), which is a part of the GU Ventures Incubator Program, all unrelated to the work presented in this paper. S.O.M. reports receiving speaker fees from Medtronic, Novartis, Novo Nordisk, Pfizer, Bayer and advisory board fees from Boehringer Ingelheim. All other authors declare no competing interests.